PROJECT SUMMARY Genome wide association studies of complex neuropsychiatric diseases, including schizophrenia (SCZ) and bipolar disorder (BD), have identified numerous risk loci that are mostly situated in non-coding regions, necessitating a systematic study of non-coding regulatory elements. It has also been established that SCZ risk loci are preferentially located within promoter and enhancer regulatory sequences of neurons and that they co- localize with expression Quantitative Traits Loci (eQTL), thus implicating specific genes. However, work that has been performed to-date has limited spatiotemporal resolution as: (1) only a few cortical regions have been examined, (2) the effect of 3D genome on transcriptional regulation across the lifespan has never been examined, and (3) studies have been limited to homogenate brain tissue or include only broadly defined neuronal and non-neuronal populations. To address these limitations, we will generate cell type-, brain region- and age period-specific high-dimensional data that will inform us of the effect of 3D genome on the transcriptional regulation and will link regulatory elements with specific transcripts. In Aim 1, we will examine the impact of SCZ and BD risk variants on 3D genome structure and transcriptional regulation. We will use fluorescence activated nuclei sorting to isolate glutamatergic and GABAergic neuronal as well as oligodendrocyte and astrocyte nuclei from five human cortical and subcortical regions relevant to SCZ and BD across five postnatal age periods. We will then generate cell-type specific annotations for gene expression and enhancer RNA (RNA-seq and CAGE-seq), open chromatin (ATAC-seq), insulators (CTCF ChIP-seq), active enhancers and promoters (H3K27ac and H3K4me3 ChIP-seq), and chromatin loop interactions (HiC and Capture-C). Using the resulting data, we will delineate cis transcriptional regulation associated with the 3D genome (including promoter-enhancer loopings) and uncover the functional consequences of SCZ and BD risk loci on enhancer-transcript units. In Aim 2, we will examine the impact of SCZ and BD risk variants on cell type-specific gene expression and epigenome QTLs. We will map RNAseq and ATACseq at the single cell level and will use cell type-specific markers and deconvolution approaches to the existing large scale transcriptome and epigenome datasets, from CommonMind consortium, psychENCODE and other projects, in order to generate cell type-specific expression and epigenome QTLs. We will then co-localize SCZ and BD risk loci with expression and fine map epigenome QTLs to define disease-associated enhancer-transcript units. Finally, in Aim 3, we will validate disease-associated enhancer-transcript units by epigenomic editing of risk loci in iPCS-derived cells. We will apply the CRISPR/Cas9 to activate (p300) or inhibit (KRAB) enhancers of the disease-associated enhancer-transcript units (Aims1-2). Lastly, we will introduce epigenomic perturbations and characterize gene expression, chromatin accessibility and chromatin loop interactions in hiPSC-derived cells. It is our expectation that these integrated analyses will enable us to assign specific regulatory units within SCZ and BD risk haplotypes to specific cell types, brain regions and age windows, thereby providing insight into the mechanisms of genetic risk for SCZ and BD.